better-colors
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze project-specific data, including source code literals and rendered UI elements, to identify and validate color systems. This creates a surface for indirect prompt injection where instructions hidden in the audited data could attempt to subvert the agent's analysis.
- Ingestion points: The skill directs the agent to "Collect every literal" from project files (
palette-structure.md) and to measure rendered pairs using a browser (SKILL.md,contrast.md). - Boundary markers: The instructions do not define delimiters or explicit "ignore instructions" prompts to be used when the agent processes the discovered color literals or rendered content.
- Capability inventory: The agent performs file searches, static analysis of code, and utilizes browser tools to measure rendered colors and contrast.
- Sanitization: There are no defined processes for sanitizing, escaping, or validating the ingested content before the agent evaluates it against the color system principles.
Audit Metadata